Re: Proposal: Generalize the RandomGen and Random classes

On Wed, Oct 6, 2010 at 7:10 PM, Antoine Latter wrote:

All of this is predicated on specializing the random-data-consumer to the random-data-producer being a significant performance win.

I am very skeptical that this is the case. I'd like to see a demonstration that specializing an RNG in this way is helpful (and that implementing / sensibly defaulting the missing methods actively gets in the way of this specialization). -Jan-Willem Maessen